21.4 brief comms MH
14/4/05
4:37 pm
Page 973
brief communications
Arboreal ants build traps to capture prey
Tiny ants construct an elaborate ambush to immobilize and kill much larger insects.
o meet their need for nitrogen in the
restricted foraging environment provided by their host plants, some arboreal ants deploy group ambush tactics in
order to capture flying and jumping prey
that might otherwise escape1–4. Here we
show that the ant Allomerus decemarticulatus uses hair from the host plant’s stem,
which it cuts and binds together with a purpose-grown fungal mycelium, to build a
spongy ‘galleried’ platform for trapping
much larger insects. Ants beneath the platform reach through the holes and immobilize the prey, which is then stretched,
transported and carved up by a swarm of
nestmates. To our knowledge, the collective
creation of a trap as a predatory strategy has
not been described before in ants.
Allomerus decemarticulatus (Myrmicinae)
is specifically associated with the Amazonian
ant-plant Hirtella physophora (Chrysobalanaceae), which houses colonies in leaf
pouches. Workers build galleried structures
on their host plant’s stems in which they pierce
numerous holes that are slightly larger in
diameter than their heads, allowing them to
enter and exit5 (Fig. 1a). First, they cut plant
hairs (trichomes) along the stems, clearing a
path. Then, using uncut trichomes as pillars,
they build the gallery’s vault by binding cut
trichomes together with a compound that
they regurgitate (for details, see supplementary information). Later, this structure is
reinforced by the mycelium of a complex of
sooty-mould species that has been manipulated by the ants.Fungal growth starts around
the holes and then spreads rapidly to the rest
of the structure.
We noted that the stems of 34 young
seedlings, which had not yet developed leaf
pouches, did not bear fungus; nine saplings
raised in a greenhouse in the absence of
Allomerus developed leaf pouches but never
bore fungus. However, 15 saplings raised in
the presence of ants bore mycelia, whose
development was limited to the galleries.
When we eliminated the associated ants
from five of the 15,the fungus on the galleries
grew into a disorganized structure, and none
of the nine new stems that developed bore
any fungus at all.
Because prey seemed to be immobilized
on the top surface of these galleries, we investigated whether these structures could be
acting as a trap. Our observations revealed
that Allomerus workers hide in the galleries
with their heads just under the holes,
mandibles wide open, seemingly waiting for
an insect to land. To kill the insect, they grasp
its free legs, antennae or wings, and move in
and out of holes in opposite directions until
T
31062 Toulouse, France
e-mail : [email protected]
†Department of Animal Biology, University of
Illinois at Urbana–Champaign, Urbana,
Illinois 61801, USA
‡Laboratoire de Psychologie Sociale de la Cognition
(ESA CNRS 6024), Université Blaise Pascal,
63037 Clermont-Ferrand, France
1. Heil, M. & McKey, D. Annu. Rev. Ecol. Evol. Syst. 34, 425–553
(2003).
2. Hölldobler, B. & Wilson E. O. The Ants (Harvard Univ. Press,
Cambridge, Massachusetts, 1990).
3. Dejean, A. & Corbara, B. in Arthropods of Tropical Forests.
Spatio-temporal Dynamics and Resource Use in the Canopy
(eds Basset, Y., Novotny, V., Miller, S. E. & Kitching, R. L.)
341–347 (Cambridge Univ. Press, Cambridge, UK, 2003).
4. Morais, H. C. Insectes Soc. 41, 339–342 (1994).
5. Benson, W. W. in Amazonia (eds Prance, G. & Lovejoy, T. E.)
239–266 (Pergamon, New York, 1985).
6. Offenberg, J. Behav. Ecol. Sociobiol. 49, 304–310 (2001).
7. van Borm, S., Buschinger, A., Boomsma, J. J. & Billen, J.
Proc. R. Soc. Lond. B 269, 2023–2027 (2002).
8. Davidson, D. W. Biol. J. Linn. Soc. 61, 153–181 (1997).
Supplementary information accompanies this communication on
Nature’s website.
Competing financial interests: declared none.
a
b
c
Predation
Prey plumage adaptation
against falcon attack
d
Figure 1 Galleried platforms built by Allomerus workers act as
traps for prey. a, Base of the lamina of two Hirtella physophora
leaves: the pouches serve as ant nests, which are interconnected
by a gallery pierced with numerous holes. b, Start of the capture
of a locust. c, One hour later, recruited workers have left the
gallery in order to bite, sting and stretch the prey. d, Completion of
a cricket capture.
the prey is progressively stretched against the
gallery and swarms of workers can sting it
(Fig. 1b–d). The ants then slide the prey
over the top of the gallery — again moving in
and out of holes, but this time in the same
direction. They move it slowly towards a leaf
pouch, where they carve it up.
Because the requirement for protein is
crucial, some arboreal ants consume parts of
the Hemiptera that they attend6, others rely
on microsymbionts to recycle nitrogen7, and
some have a thin cuticle and non-proteinaceous venom that economizes on nitrogen
use8. Nevertheless, most of them must capture prey that land on their host plant3.Given
this constraint, the tiny A. decemarticulatus
workers have developed a tripartite association with their host plant and a fungus collectively to ambush their prey.
Alain Dejean*, Pascal Jean Solano*,
Julien Ayroles†, Bruno Corbara‡,
Jérôme Orivel*
*Laboratoire d’Évolution et Diversité Biologique
(UMR CNRS 5174), Université Toulouse III,
NATURE | VOL 434 | 21 APRIL 2005 | www.nature.com/nature
everal plumage types are found in feral
pigeons (Columba livia), but one type
imparts a clear survival advantage
during attacks by the swiftest of all predators
— the peregrine falcon (Falco peregrinus)1,2.
Here we use quantitative field observations
and experiments to demonstrate both the
selective nature of the falcon’s choice of
prey and the effect of plumage coloration
on the survival of feral pigeons. This
plumage colour is an independently heritable trait3 that is likely to be an antipredator
adaptation against high-speed attacks in
open air space.
The polymorphic feral pigeon is ideal for
investigating the effects of different plumage
combinations on predator success. One
plumage phenotype –– known as the ‘wild’
variant — is blue-grey but has a white rump
between the base of the tail and the lower
back, closely resembling the feral pigeon’s
rock-dove ancestor3,4 (Fig. 1); all other
plumage types lack this contrasting rump
patch3. We tested whether the wild coloration might impart a selective advantage
over other plumage types during high-speed
diving attacks by falcons, which often exceed
157 m s1 (ref. 2; Fig. 1a).
During a seven-year period, we recorded
1,485 attacks by five adult peregrine falcons
on free-ranging flocks of feral pigeons in
Davis, California, during their daily linear
S
973
©2005 Nature Publishing Group
21.4 brief comms MH
14/4/05
4:37 pm
Page 974
brief communications
commute,and an additional 309 attacks when
three of these falcons were juveniles.For each
attack, we noted the plumage phenotype of
the selected pigeon and whether the stoop
led to a successful capture (for methods, see
supplementary information).
To determine the natural distribution of
plumage phenotypes, we banded and recorded data from 5,235 pigeons in the study area.
Adult falcons selected plumage phenotypes
for attack in the same relative proportions
that occurred naturally (Fig. 1a, inset), except
for the wild phenotype (P0.006). Juvenile
falcons, however, attacked all plumage phenotypes in proportion to their frequency in
the population.
Comparison of adult and juvenile falcons
revealed a significant difference in the
plumage phenotypes they selected as prey
(F [5,1782]2.129, P0.05; Fig. 1b). But
there was no difference between adults and
juveniles in their captured pigeons’pattern of
plumage (F [5,585]0.39, P0.85; Fig. 1b),
suggesting that both age classes learn to avoid
the wild plumage during attacks.
Adults were significantly more adept
hunters, capturing their prey on 40% of
stoops,whereas juvenile success was only 19%
(P0.0001). Both adults and juveniles,
however, showed a significant interaction
between the phenotypes selected for attack
and those captured (adult: F[5,2075]
233.7, P0.0001; juvenile: F [5,360]2.66,
P0.02; Fig. 1b). Much of this effect was driven by the wild phenotype, which accounted
for only 2% of all captures in both age classes.
To confirm the advantage afforded by a
white rump to pigeons during a high-speed
attack, we captured 756 wild and bluebarred pigeons and switched their rumppatch feathers. We then released these
pigeons and monitored predation by three
adult peregrine falcons.After plumage transfer, we found that capture rates were reversed
(Fig. 1c). The original wild phenotype now
suffered predation at the same rate as the
unmanipulated blue-barred type, whereas
the manipulated blue-barred type now had
rates of predation as low as the unmanipulated wild type (P0.0001). This indicates that
the dorsal white rump patch is important for
the survival of feral pigeons during attacks by
peregrine falcons.
All feral pigeons perform the same evasive
roll during predation by falcons (Fig. 1a).
The protective white patch may disguise
the initiation of the pigeon’s evasive roll by
contrasting conspicuous (white patch) and
cryptic targets (grey wings and body)5.A fastflying falcon primed to a conspicuous target
centered on the roll might fail to detect the
dodge initiated by the cryptic wings as the
predator closes from behind. When pursued
by predators, schools of fish in open water
and many shorebirds also display their
dark dorsal and light ventral surfaces in a display that alternates between a cryptic and a
Figure 1 Adaptive coloration in
pigeons hunted by peregrine falcons. a, Sequence of attack by a
falcon as it closes on a pigeon:
from the bottom of its dive (top
left), the falcon arcs behind the
pigeon to begin its 20–100-m
horizontal approach; the pigeon
then takes avoiding action by
dipping one wing, rapidly rotating (arrow), and rolling out of the
path of the falcon. Inset, natural
distribution of the six pigeon
plumage phenotypes in the
study population (n5,235).
b, Percentage of wild-coloured
(shown in blue) and blue-barred
(shown in red) pigeons selected
for attack (left panel) and actually
captured (right panel) from the
natural population by adult
and juvenile falcons. (Statistics
indicate comparison of all phenotypes.) c, Percentage of wildcoloured (white rump) and
blue-barred pigeons captured in
the natural population (left panel;
n203) and after plumage
transfer (right panel; n69) in a
field experiment.
conspicuous signal6–9. The use of contrasting
patterns as an antipredator mechanism is
widespread and may represent a case of convergent evolution.
Given such a selective advantage, why is
the wild phenotype not better represented
in flocks of feral pigeons? Unlike their
monomorphic ancestor, the rock dove, feral
pigeons often choose mates with plumage
unlike their own (negative assortative mating), and this maintains plumage polymorphisms3. Also, the conspicuous white rump
patch could prove disadvantageous when the
bird is near cover,such as trees,or confronted
by other predators, particularly accipiters
and falcons that typically chase down their
prey at lower speeds.
In the eastern part of the pigeon’s native
range in Eurasia, C. livia is sympatric with a
subgenus of falcons, Hierfalco, that typically
chase their prey in level flight. Pigeons from
this region lack the white rump patch1,4. The
decline in recent decades of peregrine-falcon
populations worldwide relaxed their selective
pressure on feral pigeons, but this has now
reversed, with falcons recolonizing many of
their former haunts10. Over the study period,
the proportion of wild plumage types in our
study population increased significantly relative to blue-barred types (P0.01),in parallel
with a steady increase in predation by peregrine falcons. This suggests that, although
pigeon plumage polymorphism persists, falcon predation pressure can lead to an increase
in the relative proportion of the wild pigeon
phenotype.
Alberto Palleroni*‡, Cory T. Miller†,
Marc Hauser*, Peter Marler‡
*Primate Cognitive Neuroscience Laboratory,
Department of Psychology, Harvard University,
Cambridge, Massachusetts 02138, USA
e-mail: [email protected]
†Laboratory of Auditory Neurophysiology,
Department of Biomedical Engineering, Johns
Hopkins University School of Medicine, Baltimore,
Maryland 21205, USA
‡Laboratory of Animal Communication, University
of California, Davis, California 95616, USA
1. Cade, T. J. The Falcons of the World (Cornell Univ. Press, Ithaca,
New York, 1982).
2. Tucker, V. A. J. Exp.Biol. 201, 403–414 (1998).
3. Johnston, R. F. & Janiga, M. Feral Pigeons (Oxford Univ. Press,
Oxford, UK, 1995).
4. Goodwin, D. Pigeons and Doves of the World (Cornell Univ.
Press, Ithaca, New York, 1970).
5. Bond, A. B. & Kamil, A. C. Anim. Learn. Behav. 27, 461–471
(1999).
6. Helfman, G., Collette, B. & Facey, D. The Diversity of Fishes
(Blackwell Science, Malden, Massachusetts, 1997).
7. Tinbergen, N. The Study of Instinct (Oxford Univ. Press,
London, 1951).
8. Brooke, M. de L. Funct. Ecol. 12, 339–346 (1998).
9. Lima, S. L. Wilson Bull. 105, 1–47 (1993).
10. Cade, T., Enderson, J. H., Thelander, C. G. & White, C. M. (eds)
Peregrine Falcon Populations: Their Management and Recovery
(The Peregrine Fund, Boise, Idaho, 1988).
Supplementary information accompanies this communication on
Nature’s website.
Competing financial interests: declared none.
NATURE | VOL 434 | 21 APRIL 2005 | www.nature.com/nature
974
©2005 Nature Publishing Group